From Talking Heads to Thinking Heads:
A Research Platform for Human Communication Science

The Thinking Head Project

The Thinking Head Project, funded by a $3.4M Australian Research Council/National Health & Medical Research Council Special Initiative Thinking Systems grant for 2006-2011, consists of a 15-person research team made up of computer scientists, engineers, language technologists, cognitive scientists, and performance artists. The project is centred at MARCS Auditory Laboratories at the University of Western Sydney, and Macquarie University, Flinders University, University of Canberra, Carnegie Mellon University, the Technical University of Denmark, and Berlin University of Technology. The project draws on the expertise, resources and methodological approaches of researchers in HCSNet, the Australian Research Council Network in Human Communication Science.

Postgraduate Research Scholarships

The University of Western Sydney, Macquarie University, Flinders University and the University of Canberra are offering up to 4 Postgraduate Research Scholarships under a joint ARC/NHMRC Thinking Systems Special Initiative grant to build an Embodied Conversational Agent.

One of the PhD scholarships at the University of Canberra is for a project in the area of audiovisual automatic speech recognition. The aim of this project is to advance the state of the art in AVASR and to use a combination of visual and acoustic inputs to enable the Thinking Head to recognise the words and sentences it “hears” and “sees” when spoken to by humans.

The other PhD scholarships at the University of Canberra is in the area of detecting speaker characteristics such as gender, age, accent and emotional state from the visual and acoustic inputs of stereo cameras and a microphone array, so that the Thinking Head can respond appropriately and differently to a school child, elderly person, person with a foreign accent, or someone who is curious, bored, frustrated or angry.

The projects will be carried out within the Human-Computer Communication Lab in the School of Information Sciences and Engineering and in the National Centre for Biometric Studies of the University of Canberra.

If you are considering applying for one of these scholarships, please contact Michael.Wagner@canberra.edu.au

Scholarships are also available in other areas of the project at other participating laboratories:

• Integration of Auditory and Visual Speech Input and Output at Flinders University; and 
• Automatic Audio-Visual Speech Recognition and Detection of Audio-Visual Speaker Characteristics at the Macquarie University.